How can I identify a file type from a blob/filestream?

Question

We bought an "off the shelf" application a lonnng time ago that is capable of storing files as a blob within SQL Server. We've noticed that the database has more than doubled in size within the past six months due to more frequent usage of this blob field by one department. As a result, the application has become painfully slow.

I've been tasked with removing the blob field from the database and saving the file onto the actual file system. Unfortunately, the application does not store what the file type is within the database. Although I can read the file as it exists in the database, I don't know what extension to save the file as. The application's support desk no longer supports this version of the software and will not talk to us about extracting the data. Unfortunately, we do not have access to their source code.

Any suggestions would be greatly appreciated! Thanks in advance!

Answer 1

you can look at the first few bytes and figure it out for the most common file types

http://www.garykessler.net/library/file_sigs.html

Answer 2

Don't save it as any type. Save it as a file with no extension. If you don't know what it is, don't fake it. If the app that saved it requests it, return it from the filesystem the same way it would be returned from the database; as binary data. The database doesn't care what type of data the Binary Object is; neither should you.

Answer 3

You might try using TriD http://mark0.net/soft-trid-e.html

It will scan the files and try to determine the extension.

Answer 4

You could use the FindMimeFromData() function in UrlMon.dll (using pinvoke).

See this page for an example and this MSDN page for the documentation of the function.